Artifact Explanation on SynthScars (test)
24.7ROUGE ScoreLEGION
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LEGIONTraining Split=SynthScars training dataset2026.02 | 24.7 | 58.9 | |
| Qwen2.5-VL-7B + ArtiAgentFine-tuning=100K training set generated by ArtiAgent2026.02 | 19.6 | 57.8 | |
| InternVL3.5-8B + ArtiAgentFine-tuning=100K training set generated by ArtiAgent2026.02 | 17.9 | 51.3 | |
| GPT-4o2026.02 | 12.5 | 40.4 | |
| GPT-52026.02 | 12 | 46.1 | |
| Qwen2.5-VL-7BFine-tuning=Vanilla2026.02 | 11.5 | 36.2 | |
| Gemini-2.5-Pro2026.02 | 10.3 | 47.4 | |
| InternVL3.5-8BFine-tuning=Vanilla2026.02 | 5 | 18 |